> ## Documentation Index
> Fetch the complete documentation index at: https://docs.ai-stats.phaseo.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Streaming

> Use server-sent events (SSE) for low-latency token delivery.

Streaming returns incremental output as SSE frames so users see responses earlier.

## Supported endpoints

* `/v1/responses`
* `/v1/chat/completions`
* `/v1/messages`

## Request

```bash theme={null}
curl https://api.phaseo.app/v1/responses \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-5-nano",
    "input": "Stream a short hello message.",
    "stream": true
  }'
```

## Response

```text theme={null}
data: {"id":"resp_...","object":"response","status":"in_progress",...}
data: {"type":"response.output_text.delta","delta":"Hello"}
data: {"type":"response.completed",...}
data: [DONE]
```

## SDK options

* TypeScript SDK: `client.streamResponse()` and `client.streamText()`
* Python SDK: `client.stream_response()` and `client.stream_text()`
* Vercel AI SDK: `streamText(...)` with `@ai-stats/ai-sdk-provider`

## Important limitation

At current gateway request-validation level, combining tools with `stream: true` is rejected (`400 invalid_request`). Use non-streaming for tool-calling loops.

## Next guides

1. [SSE Event Reference](./streaming-sse)
2. [Streaming in Production](./streaming-production)